Pre Touch

Pre-touch sensing focuses on equipping robots with the ability to perceive objects' properties and poses *before* physical contact, enhancing manipulation capabilities. Current research emphasizes developing advanced sensors integrating techniques like ultrasound, optoacoustics, and spectroscopy to gather information on material composition, distance, and object geometry. Deep learning models are increasingly used to process sensor data efficiently, focusing sensing on key object features to improve pose estimation accuracy. This technology promises significant improvements in robotic dexterity and adaptability, particularly in handling delicate objects or in situations where precise manipulation is crucial.